Meta is a global technology company that helps people connect, find communities, and grow businesses. They’re best known for their social media platforms Instagram and Facebook, but also offer Whatsapp, Messenger, Workplace, Oculus, Portal, and Novi.
I was on the Growth Ecosystems team, which is concerned with the maintenance and growth of users on the Facebook app. It’s a high-level ecosystem team, which means that we weren’t concerned with any one product but instead spanned all of them in our reporting, analysis, and forecasting.
My internship focused on marginal users (low-engaged users) on Facebook. Right now, the Growth Ecosystems team tracks Monthly Active Users (MAU) and Daily Active Users (DAU).
I first algorithmically determined a threshold for user activity levels to define marginal users to be tracked alongside MAU and DAU. Then I analyzed the trends and activity of those users globally. Ultimately, I leveraged statistics, visualization, linear regression, and clustering to uncover key regional and global insights about marginal users.
There were two components of my analysis that were generalizable for use by other teams: the first was a general user activity clustering guide that became one of the first items in a repository of reproducible analyses for data scientists company-wide, and the second was product diversity, a novel method of measuring and quantifying user activity.
My project has four main components: Marginal User Definition, Regional User Activity Clustering, a Reproducible Clustering Guide, and Product Diversity.
Purpose: Determine a threshold for defining a set of low-engaged users to be tracked alongside MAU and DAU.
<aside> 💡 Impact: Standardized the active days threshold for low-engaged users and defined a new user type to help measure and evaluate user growth and health.
</aside>